While there are some interesting conversations about numpy.argmin() function’s performance (e.g. here, here and here), I want to stay away from making any comments. I also do not claim that my alternative is necessarily faster or slower than np.argmin() or np.argmax(), an alternative is an alternative.
| >>> import numpy as np | |
| >>> x = np.arange(127, 255, 50) | |
| >>> x | |
| array([127, 177, 227]) | |
| >>> x.argmin() | |
| 0 | |
| >>> x = x[::-1] | |
| >>> x | |
| array([227, 177, 127]) | |
| >>> x.argmin() | |
| 2 | |
| >>> | |
| >>> # an alternative ... | |
| ... | |
| >>> x = np.arange(127, 255, 50) | |
| >>> n = range(len(x)) | |
| >>> n[x.tolist().index(min(x))] | |
| 0 | |
| >>> x = x[::-1] | |
| >>> n[x.tolist().index(min(x))] | |
| 2 | |
| >>> |
